AI Data Management: A Key Compliance Challenge for Insurers
Insurers face significant hurdles in meeting regulatory requirements for artificial intelligence (AI) data management. A major issue is the limited understanding among both insurance companies and regulators about the data science behind AI models. Scott Harrison, co-founder of the American InsurTech Council, highlights that many insurers struggle to clearly explain where AI data originates and how their models function to the satisfaction of regulators.
Regulatory bodies demand transparency. While insurers recognize the importance of protecting consumers, there is concern that overly strict regulations could hinder innovation and favor only the largest, well-funded players in the market.
Emerging AI Regulations in Insurance
AI regulation within insurance is relatively new. The National Association of Insurance Commissioners (NAIC) introduced regulatory guidance in December 2023, which has already been adopted by 24 U.S. states. Internationally, the European Union implemented its AI Act in May 2024, and the U.K. followed with principles-based AI regulation in February 2024.
Parul Kaul-Green, CEO of Eudaimon Consulting and former executive at Liberty Mutual and AXA XL, explains that the EU AI Act will influence global standards. This “Brussels effect” means multinational insurers operating across borders will likely have to align with the EU’s comprehensive, risk-based framework.
Legal Cases May Shape Future AI Oversight
Kaul-Green notes that outside the EU, regulators rely on existing laws addressing bias, discrimination, data privacy, and unfair trade practices to manage AI risks. However, AI introduces new business methods that may require additional legal clarity.
She suggests that future court cases could supplement current regulations by addressing AI-specific issues like discriminatory practices or consumer protection breaches not yet fully covered by existing laws.
Practical Compliance Tips for Insurers
Many insurers use AI to digitize processes that were previously manual. When regulators investigate these changes, insurers must provide clear explanations for any decisions driven by AI. Harrison warns that simply stating “the algorithm made the decision” will not suffice, especially in cases like claims denials.
Francesca Blythe, partner at Sidley Austin LLP, advises insurers to begin implementing AI compliance frameworks even if they’re not perfect. She stresses starting small and building on existing measures rather than waiting for an ideal solution.
- Develop clear documentation on AI data sources and model functionality.
- Maintain transparency to satisfy regulatory inquiries.
- Implement incremental compliance measures and improve over time.
- Stay informed about evolving international AI regulations.
For insurance professionals looking to strengthen their AI knowledge and compliance skills, exploring relevant AI training can be valuable. Resources like Complete AI Training’s courses tailored for insurance professionals offer practical insights into AI implementation and governance.
Your membership also unlocks:
